AWS Transcribe Parser

AWS Transcribe Parser

I was listening to a podcast Game Dev Loadout and really wanted to find out some specific software that one of the devlopers was using but the podcaster had not put out any transcript yet. These episodes are usually 45 minutes to an hour in length and having to relisten to the entire episode just to hear the one mention of a software was a bit much. This is where the AWS Transcribe service came in. I get 60 minutes of transcribe free within the free tier and was able to process the single episode with ease.

The output from AWS Transcribe is in JSON and not easily human readable. I could have searched the document for what I thought the name was, but this was going to be hit or miss depending on my spelling. What I really needed was a readable format in paragraphs so I could quickly scan the converstation and find the name I was looking for. I went about parsign the JSON and outputting this into a nicer format so I could scan the resulting HTML document. I then realized that I could expand this further to the point of a live edit document that others could include their transcribe file and output to HTML so the result could be dropped into their blog.

Here is a sample of the JSON output from the AWS Transcribe service. While amazing amount of details are contained in the output, it can be difficult for someone that would just like the output to be in a format that can be dropped into blog post.

Here is the format from the tool. With the help of some open source libraries, I was able to make an interactive editing experience from the JSON AWS Transcribe output.

Give it a try yourself by using the published version. Demo Link: http://wvprogramming.com/transcribe/parser.html

If you would like to take a look at the code or fork a version off for yourself, here is the GitHub repo. GitHub Link: https://github.com/curtisdarst/awsTranscribeHTMLFormatter